Anatomical Mapping for Atrial Fibrillation Ablation: A Head-to-Head Comparison of Ultrasound-Assisted Reconstruction versus Fast Anatomical Mapping ROBERTO RORDORF, M.D.,* ENRICO CHIEFFO, M.D.,* SIMONE SAVASTANO, M.D.,* ALESSANDRO VICENTINI, M.D.,* BARBARA PETRACCI, M.D.,* VALENTINA DE REGIBUS, M.D.,* ADELE VALENTINI, M.D.,† CATHERINE KLERSY, M.D.,‡ ROBERTO DORE, M.D.,† and MAURIZIO LANDOLINA, M.D.* From the *Department of Cardiology, IRCCS Fondazione Policlinico S. Matteo, Pavia, Italy; †Department of Radiology, IRCCS Fondazione Policlinico S. Matteo, Pavia, Italy; and ‡Biometry & Statistics, IRCCS Fondazione Policlinico S. Matteo, Pavia, Italy
Background: Accuracy in left atrial (LA) anatomical reconstruction is crucial to the safe and effective performance of catheter ablation of atrial fibrillation (AF). The aim of this study was to evaluate the accuracy of LA reconstruction performed with intracardiac echocardiography (ICE) as compared to fast anatomical mapping (FAM) both integrated in the CARTO mapping system (Biosense Webster, Diamond Bar, CA, USA). Methods: A multislice computed tomography (MSCT) was preacquired from 29 patients with AF who underwent catheter ablation and 3D-LA geometry was reconstructed using both ICE and FAM separately. The accuracy of the LA anatomical definition was evaluated by comparing LA volumes, LA and pulmonary vein (PV) diameters obtained using ICE and FAM versus MSCT (gold standard). Results: Anterior-posterior and superior-inferior LA diameters were shorter in ICE versus MSCT (32 ± 10 vs 46 ± 9 mm and 48 ± 7 vs 53 ± 7 mm, P < 0.01) but similar in FAM versus MSCT (45 ± 9 vs 46 ± 9 mm and 52 ± 10 vs 53 ± 7 mm). Latero-septal LA diameter was similar in ICE versus MSCT (63 ± 11 vs 63 ± 9 mm) but larger in FAM versus MSCT (69 ± 9 vs 63 ± 9 mm, P < 0.001). LA volume was lower in ICE versus MSCT (73 ± 30 mL vs 116 ± 45 mL, P < 0.0001) and slightly larger in FAM versus MSCT (132 ± 45 vs 116 ± 45 mL, P = 0.06). PV diameters were similar in FAM versus MSCT but significantly underestimated with ICE. Conclusions: Overall accuracy in the LA and PV anatomical reconstruction was found to be superior with FAM compared to ICE-guided approach. ICE resulted in a significant underestimate of both LA and PV dimensions, while FAM slightly overestimated LA geometry. (PACE 2015; 38:187–195) atrial fibrillation ablation, intracardiac echocardiography, fast anatomical mapping, CARTO system
Introduction Left atrial (LA) three-dimensional (3D) reconstruction is essential for the performance of safe and effective radiofrequency (RF) catheter ablation of atrial fibrillation (AF). Different techniques have been proposed for recreating LA anatomy in order to guide AF ablation: point-by-point electroanatomical mapping,1 intracardiac echocardiography (ICE),2,3 and fast anatomical mapping (FAM). Specifically, these two latter technologies have recently been developed and integrated
Address for reprints: Roberto Rordorf, M.D., Department of Cardiology, IRCCS Fondazione Policlinico S. Matteo, P.le Golgi 19, 27100 Pavia, Italy. Fax: 39-0382-503161; e-mail: [email protected] Received February 25, 2014; revised September 21, 2014; accepted September 24, 2014. doi: 10.1111/pace.12539
R R into the CARTO 3 System with CARTOMERGE Module (Biosense Webster, Diamond Bar, CA, USA). The ultrasound-based 3D imaging system R (CARTOSOUND Module, Biosense Webster)4,5 allows to recreate LA and pulmonary vein (PV) derived from 2D ICE images obtained using an ICE catheter placed in the right atrium (RA). FAM allows a quick recreation of 3D chamber geometries by moving sensor-based catheters placed in the left atrium. Both these methods can be used alone or after integration with a preacquired computer tomography (CT) or magnetic resonance (MR) image of the LA and PV and have the advantage of limiting x-ray exposure. Nowadays, few data comparing these two real-time reconstruction strategies are available.6 The primary aim of this study was to compare the accuracy in overall reconstruction of LA and PV geometries obtained with ICE versus FAM, considering CT images as the gold standard. A mapping technique that
©2014 Wiley Periodicals, Inc. PACE, Vol. 38
February 2015
187
RORDORF, ET AL.
Table I. Baseline Characteristics of the Study Population Age (years) Gender (M/F) Smoking Systemic hypertension Diabetes Dyslipidemia AF (paroxysmal/persistent) AF onset-AF ablation time (months) Number of antiarrhythmic drugs before ablation 1 drug 2 drugs 3 drugs LV ejection fraction (< 50%) Left atrial diameter (mm)
56 ± 9 28/1 14 (48%) 17 (58%) 2 (10%) 12 (41%) 16/13 88 ± 74 7 (24%) 9 (31%) 13 (45%) 6 (21%) 45 ± 6
AF = atrial fibrillation; LV = left ventricular.
allows to recreate a precise anatomy of both the LA and PV could potentially be used as a stand-alone tool.7 Therefore the primary objective of our study tried to address the issue whether the mapping techniques under study could be used without the need for preacquired CT or MR images. The secondary objective of the study was to evaluate the accuracy of ICE versus FAM in the integration process with CT images. Methods Study Population Twenty-nine consecutive patients (28 males, age 56 ± 9 years) with symptomatic drugrefractory AF underwent RF catheter ablation at the Electrophysiologic Unit of the Department of Cardiology and were enrolled in the study. The patient population had paroxysmal or persistent AF refractory to at least one antiarrhythmic drug. Before the procedure both transthoracic and transesophageal echocardiographic evaluations were performed. LA dimension and left ventricular ejection fraction were estimated and the presence of thrombi within the cardiac chambers was excluded. Oral anticoagulation was interrupted 2 days before hospital admittance and intravenous heparin was administered during the procedure to maintain activated clotting time > 300 ms.8 The study protocol was approved by the institutional ethical committee and written informed consent was obtained from all patients. Baseline characteristics of the study population are summarized in Table I. Multislice Computed Tomography All patients underwent multislice CT (MSCT) to reconstruct PV and LA anatomy. MSCT was
188
obtained on a Siemens Somatom Definition with 64-slice dual-source configuration (Siemens AG, Healthcare Sector, Erlangen, Germany) within 24 hours before the ablation procedure to avoid significant modifications in LA anatomy and volume secondary to changes in preload and afterload conditions. Retrospectively electrocardiogram-gated spiral scanning was performed with a collimation of 1.2 mm during injection of 80 mL of contrast medium (Iomeron 400 mg I/mL, Bracco, Milan, Italy) at a flow rate of 4 mL/s. Cardiac telediastolic phase was automatically determined in patients with sinus rhythm, while manual editing was performed on patients with AF to minimize the motion artifact. Images were acquired during the endexpiratory phase. Multiplanar reconstruction, maximum intensity projection, and volume rendering were used to define and describe LA anatomy. LA volume was measured with a dedicated volume analysis software (Syngo VE32B Volume, software version 1.0, 2008, Siemens AG). The LA was manually outlined on axial images by using a 3-mm thickness sequence; afterwards the contours were semiautomatically determined and volume was calculated. Ostial PV diameters, LA anterior-posterior, lateralseptal, and superior-inferior diameters were also manually determined. MSCT images (slice thickness 1.5 mm, position increment 0.7 mm) were then stored on a CD and data were loaded into the image integration module of the R mapping system (CARTOMERGE Module, Biosense Webster). Afterwards a segmentation process of MSCT images was performed as previously described.9,10 3D reconstructions of LA and PVs were isolated from the surrounding structures and exported to the mapping system for the subsequent mapping and integration processes. LA and PV anatomy obtained with MSCT was considered the gold standard and used to evaluate the accuracy of anatomical reconstruction obtained with ICE and FAM. Anatomical Reconstruction with ICE After vascular access was obtained, a 10-Fr R ICE catheter (SOUNDSTAR Catheter, 64-element, 5.5–10.0 MHz, Biosense Webster) was placed in the left femoral vein and advanced into the RA. The ICE catheter tip incorporates a sensor to establish its location and direction and to show the ultrasound beam projection in the mapping system. Sequential clockwise manipulation of the ICE probe from the RA allowed for visualization of the LA, LA appendage (LAA), and PVs. 2D ultrasound segment images were acquired during
February 2015
PACE, Vol. 38
ANATOMICAL MAPPING FOR ATRIAL FIBRILLATION ABLATION
end-expiratory phase and gated to coronary sinus atrial electrogram in patients in sinus rhythm and to 80% of the R-R interval in patients with AF. ICE was performed using a Sequoia Cypress ultrasound system (Siemens Medical Solutions USA, Mountain View, CA, USA), which transferred real-time ICE data to the mapping system. The endocardial surface contours of each ultrasound image were manually drawn. All contours were used to automatically generate a single 3D anatomical map that included both LA R and PVs with the use of the CARTOSOUND Module (Biosense Webster). The volume of the chambers R was automatically provided by the CARTOSOUND Module as volume subtended to the interpolated map surface. Anatomical Reconstruction with FAM After double-transseptal catheterization, a 3.5-mm cool saline-irrigated ablation catheter R R (NAVISTAR THERMOCOOL Catheter, Biosense Webster) and a multipolar circular catheter R (LASSO NAV Catheter, Biosense Webster), both R compatible with CARTO mapping system, were advanced in the LA and PVs and used for anatomical reconstruction. High-resolution 3D chamber volumes were created from continuous recording of the catheter movements within the cardiac chambers to outline the structures. Anatomical rendering was obtained by the mapping system in a stable mode calculating the average of the catheter positions in the running second. The mapping catheter was placed 1 cm inside each PV and multiple positions were acquired while the catheter was rotated inside the vein and pulled back toward the ostium. The highest definition setting was used to better separate each vein and the lateral veins from the LAA. A 3D shell of the LA and LAA was created roving R R R the NAVISTAR , THERMOCOOL , or LASSO NAV catheters into the atrium: a lower definition setting was applied for the LA to obtain sufficiently fast and detailed atrial reconstruction. During volume sampling, catheter movements during inspiration phase were manually excluded. The R CARTO system automatically calculates the volume of the chambers as the volume subtended to the interpolated map surface. Electroanatomical points were not acquired and adapted to LA surface during PV and LA reconstruction with FAM. Anatomical reconstruction of the LA was performed both with ICE and FAM during the same cardiac rhythm (either sinus rhythm or AF) as during the acquisition of MSCT images in all patients. Anatomical mapping was performed in sinus rhythm in 16 patients and in AF in 13 patients.
PACE, Vol. 38
Accuracy of Overall Anatomical Reconstruction of LA and PVs In order to establish the overall accuracy of 3D anatomical reconstruction of LA and PVs, obtained with ICE versus FAM, MSCT images were considered the gold standard. PV, LA diameters, and LA volumes were measured with ICE and FAM and compared to those obtained with MSCT. The anterior-posterior, latero-septal, and superiorinferior diameters of the LA were assessed with MSCT as in Figure 1 and compared to those obtained with both ICE and FAM reconstruction. Figure 2 shows an example of the superior-inferior LA diameter calculated with MSCT, FAM, and ICE in the same patient. In addition, the overall LA volume resulting from ICE and FAM was compared to that obtained with MSCT. In the computation of the overall LA volume, care was taken to include the proximal part (1 cm inside) of each PV. The ability of each mapping technique to identify the number of PVs per patient and the presence of separate versus common ostia were also assessed. Finally, PV diameters were measured with each technique. PV diameters on ICE were measured at the widest point; PV diameters on MSCT and FAM were measured at the same angulation as with ICE to obtain comparable data. LA and PV diameters were measured by two operators independently and interoperator variability was assessed. The time needed for LA reconstruction with ICE versus FAM was calculated and x-ray exposure time was also analyzed.
Accuracy of the Registration Process with MSCT After creating an anatomical reconstruction with ICE and FAM, each map was integrated with MSCT separately. First of all, the ICE anatomical map was displayed adjacent to the MSCT image. A landmark point was placed at a distinct anatomical structure (usually at the PV ostia) on the ICE map and at the corresponding point on the MSCT image. Visual alignment and surface registration was then performed as described previously.4 The FAM map was then displayed adjacent to the MSCT image and the same integration process was performed for the FAM anatomical map, without considering the ICE contours. A new landmark point was identified on the FAM map and on the corresponding point on the MSCT, and a new visual alignment and surface registration process were performed. The accuracy of the registration process was assessed by evaluating the mean distance and the standard deviation of the distances between the MSCT surface and each point on the ICE and FAM map separately.
February 2015
189
RORDORF, ET AL.
Figure 1. Left atrial diameters. Left panel: The anterior-posterior diameter (sagittal diameter) of the left atrium calculated as the distance between the central part of the posterior wall to the central part of the mitral annulus. Middle panel: The superior-inferior diameter (longitudinal diameter) calculated as the distance between the central part of the roof to the central part of the inferior wall. Right panel: The latero-septal diameter (transverse diameter) calculated as the distance between the central part of the lateral carina to the central part of the septal carina.
Figure 2. Right lateral view showing LA superior-inferior diameter calculated with FAM (left), MSCT (middle), and ICE (right) in the same patient. FAM = fast anatomical mapping; ICE = intracardiac echocardiography; LA = left atrial; MSCT = multislice computed tomography.
Statistical Analysis Data are presented as mean ± standard deviation for continuous variables and as count and percentage for categorical variables. Statistical comparisons were performed with the two-tailed paired Student’s t-test. Bland and Altman limits of agreement and Lin’s concordance correlation coefficient (Lin R), with 95% confidence intervals (95% confidence interval) were computed to assess interoperators and intermethods agreement. P < 0.05 was considered statistically significant. The Bonferroni correction was applied when comparing each of the two methods against MSCT (P < 0.025 for significance). Statistical analysis was performed with MedCalc (MedCalc Software,
190
Ostend, Belgium) and Stata 13 (StataCorp, College Station, TX, USA). Results The time required for a comprehensive LA and PV anatomical reconstruction was similar for ICE and FAM techniques (33 ± 14 minutes vs 30 ± 12 minutes, P = 0.410). Fluoroscopy time needed for ICE reconstruction was significantly lower than the time required for FAM reconstruction (2.1 ± 4.4 minutes vs 8.9 ± 4.6 minutes, P < 0.0001). A mean of 23 ± 10 ultrasound contours were used to obtain a global anatomical rendering of the LA and PVs. There were no major complications related to the ablation procedure or to the protocol.
February 2015
PACE, Vol. 38
ANATOMICAL MAPPING FOR ATRIAL FIBRILLATION ABLATION
Accuracy of Overall LA and PV Anatomical Reconstruction The ability of ICE anatomical mapping to recreate a comprehensive LA 3D shell resulted suboptimal due mainly to an underestimate of the LA dimension in the sagittal and longitudinal diameter. Indeed, compared to the gold-standard MSCT, ICE-derived anteriorposterior and superior-inferior diameters resulted significantly shorter (32 ± 10 mm vs 46 ± 9 mm, P < 0.0001, 48 ± 7 mm vs 53 ± 7 mm, P = 0.001). Conversely, the latero-septal LA diameter obtained with ICE reconstruction was comparable to MSCT (63 ± 11 mm vs 63 ± 9 mm, P = 0.956). On the contrary, FAM-derived 3D LA reconstruction was more similar to that obtained with MSCT except for a significant overestimation of the LA dimension in the transverse diameter. The latero-septal LA diameter was significantly larger with FAM compared to MSCT (69 ± 9 mm vs 63 ± 9 mm, P ࣘ 0.001). Both anterior-posterior and superior-inferior LA diameters were similar in FAM compared to MSCT (45 ± 9 mm vs 46 ± 9 mm, P = 0.107, 52 ± 10 mm vs 53 ± 7 mm, P = 0.511). Figure 3 summarizes the LA diameters obtained with ICE and FAM compared to MSCT. Similar results were found when the analysis was repeated separately in patients performing anatomical mapping in sinus rhythm and AF (see Table II). Overall LA volume was significantly underestimated by ICE compared to MSCT (73 ± 30 mL vs 116 ± 45 mL, P < 0.0001) and slightly overestimated by FAM versus MSCT (132 ± 45 mL vs 116 ± 45 mL, P = 0.06). Regarding the accuracy of ICE 3D reconstruction, the analysis was repeated after excluding those patients with less than 20 ICE contours available. Similar to the general population, ICE-derived anteriorposterior, superior-inferior diameters, and overall LA volume were significantly shorter compared to MSCT (32 ± 9 mm vs 46 ± 10 mm, P < 0.0001, 48 ± 7 mm vs 53 ± 10 mm, P = 0.001, 77 ± 31 mL vs 120 ± 47 mL, P < 0.0001). All patients showed four main PVs. FAM allowed to correctly identify all PVs in the entire population. On the contrary, ICE did not correctly reconstruct all PVs in five patients: in one patient a right inferior PV was not identified and in four patients the quality of the anatomical rendering was considered unsatisfactory (three right veins and one left vein). In 24 of 29 patients, the four PVs showed separated ostia. In five patients a common ostium was detectable (four left and one right). FAM allowed to identify a common ostium in all cases while in one case ICE identified
PACE, Vol. 38
the two veins but not the proximal part of the common ostium. Diameters of all four PVs were underestimated with ICE versus MSCT. On the contrary, PV diameters obtained with FAM were comparable to MSCT. Figure 4 summarizes the PV diameters obtained with the different mapping techniques. PV diameters obtained with ICE were still shorter compared to MSCT also when the analysis was limited to those patients with 20 or more acquired ICE contours. A better accuracy in overall anatomical reconstruction with FAM versus ICE was also confirmed by intermethods agreement analysis. Indeed, overall agreement with preacquired MSCT was superior with FAM for both LA and PV diameters (see Fig. S1). Interoperator variability was excellent for PV and LA diameters measurements with all methods (see Fig. S2). Accuracy of the Registration Process of ICE and FAM Maps with MSCT To enable a registration process with preacquired MSCT images, a fiducial landmark point was identified on each recreated anatomical map. A landmark point at the PV ostia was used in the majority of the cases in both FAM and ICE reconstructions. Specifically, a landmark point was identified at a PV ostium in 19 cases and at the PV carina in 10 cases with FAM. Similarly, with ICE a landmark point was identified at a PV ostium in 19 cases, at the PV carina in six cases, and within the LA in four cases. After each separate registration process, the accuracy of the map to MSCT merge was superior with ICE compared to FAM. Actually, both the mean distance and the standard deviation of the distances between the MSCT surface and each point on the reconstructed anatomical map were significantly lower for ICE compared to FAM (2.5 ± 0.5 mm vs 3.6 ± 0.7 mm; 1.8 ± 0.4 mm vs 2.9 ± 0.8 mm, P < 0.0001 for both). Displacement of the landmark point from its initial location after the merge process was similar for ICE versus FAM (8 ± 5 mm vs 8 ± 4 mm, P = 0.914). Discussion Our study reports a head-to-head comparison of two techniques for LA and PV anatomical reconstruction to guide AF ablation and highlights important differences between ICE-guided as opposed to FAM 3D rendering, with significant clinical implications for physicians performing ablation. We found that 3D ultrasound-derived anatomical reconstruction yields a significant underestimate of true LA volume, in agreement with
February 2015
191
RORDORF, ET AL.
Figure 3. Comparison of LA diameters obtained with ICE and FAM versus MSCT. AP = anterior-posterior; SI = superior-inferior; LS = latero-septal; ICE = intracardiac echocardiography, MSCT = multislice computer tomography; FAM = fast anatomical mapping. Table II. Left Atrial Diameters in Patients Performing Anatomical Mapping in Sinus Rhythm and in Atrial Fibrillation Anterior-Posterior Diameter (mm)
Patients in SR (n = 16) Patients in AF (n = 13)
Superior-Inferior Diameter (mm)
Latero-Septal Diameter (mm)
ICE
MSCT
FAM
ICE
MSCT
FAM
ICE
MSCT
FAM
32 ± 9* 30 ± 12*
44 ± 8 50 ± 9
43 ± 8 47 ± 9
48 ± 7 48 ± 6*
52 ± 8 53 ± 5
52 ± 10 51 ± 11
61 ± 12 67 ± 7
59 ± 5 66 ± 8
68 ± 10** 71 ± 6
*P < 0.01 for ICE versus MSCT; **P < 0.01 for FAM versus MSCT.
previous data from the literature.6,7,11 However, the analysis of three different LA diameters in our study gives a better understanding of the limitations of ICE in recreating a comprehensive anatomical reconstruction as described in Figure 3. Actually, our data highlight that the underestimate of LA volume is especially due to a significant misjudgment of the true measurements of the longitudinal and sagittal diameters. The use of an ICE catheter placed in the RA allows a clear definition of the transversal extension of the LA, through a simple clockwise rotation of the catheter from lateral to more septal views. Further manipulation and angulation of the catheter allows reconstruction of a 3D rendering of the LA, but still with a certain degree of underestimate of the true LA dimensions, probably due to a more challenging definition of the anterior, superior, and inferior borders of the LA. This limitation could potentially be overcome by placing the ICE catheter in the coronary sinus, in the right ventricle or directly in the LA as previously described,3 but a more extensive approach is time consuming and could increase complications. The aim of our study was to assess the accuracy of ultrasound LA reconstruction using a simple approach and taking a reasonable time. Therefore,
192
we limited the ultrasound exploration by placing the ICE catheter in the RA only. Our study assessed the accuracy of FAM for recreating a comprehensive 3D anatomical reconstruction of the LA. We demonstrated that FAM allows generation of a 3D shell of the LA that is very similar to that obtained with MSCT. It is worth noting that FAM-derived longitudinal and sagittal diameters almost matched LA measurements resulting from MSCT. The transverse diameter was the only LA dimension that was significantly overestimated by FAM. This could be due to atrial wall deformation caused by catheter manipulation: the lateral carina is directly opposite the transseptal access and this could explain why greater forces are usually applied to the mapping catheter in this area. This limitation could be probably overcome by the use of contact force sensing catheters.12 Another possible explanation for latero-septal diameter overestimate is that the FAM algorithm could be affected by higher interpolation between adjacent veins in the carina, despite the use of a highdefinition setting in this area. A slight overestimation of LA dimensions with FAM compared to MSCT could also be explained by the fact that we used a CARTO 3
February 2015
PACE, Vol. 38
ANATOMICAL MAPPING FOR ATRIAL FIBRILLATION ABLATION
Figure 4. Comparison of PV diameters obtained with ICE and FAM versus MSCT. FAM = fast anatomical mapping; ICE = intracardiac echocardiography; LIPV = left inferior pulmonary vein; LSPV = left superior pulmonary vein; MSCT = multislice computer tomography; PV = pulmonary vein; RIPV = right inferior pulmonary vein; RSPV = right superior pulmonary vein.
software version that does not allow respiratory gating. Even better accuracy in 3D rendering of the LA is to be expected with the more recent CARTO 3 version embedded with the respiratory gating module.13 Another important finding of our study is that FAM was more reliable compared to ICE in rendering the anatomy of the PVs. In our experience FAM was able to correctly identify all PVs, including five cases with common ostia, while ICE was considered unsatisfactory in a few cases. Moreover, FAM was more consistent with MSCT in PV diameter definition. Actually, the PV diameters obtained with FAM nearly matched those derived from MSCT. On the contrary, ICE-derived PV diameters were significantly smaller compared to those obtained with CT. This is in agreement with data from other groups.4,14 Left PVs are usually visualized with ICE in longitudinal cross-section, potentially missing the middle and widest part of the vein. Regarding
PACE, Vol. 38
right-sided PVs, these are usually visualized in transversal cross-section and the underestimate of the true diameter could be due to the more challenging visualization of the exact PV-LA junction. The second aim of our study was to evaluate the accuracy of the integration (Merge) process15 with MSCT of the two mapping techniques under investigation. ICE-derived anatomy yielded a better integration with MSCT compared to FAM. This finding might appear counterintuitive to the aforementioned results of a more comprehensive overall 3D LA rendering obtained with FAM. Nevertheless, compared to FAM, ICE is likely to provide a more accurate direct visualization of the fiducial points used for a more reliable superimposition with MSCT. Although ICE was probably insufficient for correctly visualizing some areas of the anterior part of the LA, it was able to define the posterior wall and the posterior PV-LA junction very precisely to allow a good
February 2015
193
RORDORF, ET AL.
alignment with MSCT. Another possible explanation is that the FAM-derived anatomical map is made up of a greater number of “virtual points” compared to ICE, given by the continuous acquisition during catheter manipulation. The greater “virtual point” density in FAM versus ICE increases the chances of mismatch between the MSCT surface and the recreated anatomical map. Our finding of a better integration accuracy with ICE versus FAM should be, however, interpreted with caution as it is based only on the overall mean distance between the MSCT surface and the recreated anatomical maps. Nonetheless, our study corroborates previous data suggesting the utility of ICE for integration processes.4 Clinical Implications A variety of issues are taken into account by electrophysiologists when choosing the preferred mapping technique for performing AF ablation: procedure and fluoroscopy time, the time spent in the left chambers, cost efficacy, and safety. The ideal mapping technique should be quick, reduce both fluoroscopy time and the time spent in the left chamber, and provide accurate anatomical information that could potentially replace MSCT. Our data add some important insights and illustrate the advantages and drawbacks of each technique. ICE-derived mapping techniques reduce both fluoroscopy time and the time spent in the left atrium, but provide an unsatisfactory overall 3D rendering, at least in a relative short procedural time, making this technique unsuitable in our opinion for replacing MSCT. Nonetheless, the ICE direct visualization of distinct anatomical structures allows a good alignment with MSCT and makes this technique suitable for integration processes. On the other hand, in our experience FAM allowed a more accurate rendering of both
LA and PV, which could potentially replace the need for preacquired MSCT, with a reduction in both procedural expenses and radiation exposure. Limitations The data presented do not allow a comprehensive analysis of the safety and efficacy issues of ICE and FAM for performing AF ablation. Our study was limited to a head-to-head investigation of the two anatomical mapping techniques, with the advantage of performing the comparison on the same patient. In our population, we did not find supernumerary veins or LA diverticula. This is probably due to the small sample size of our study. A larger study would be required to analyze the accuracy of ultrasound and FAM mapping in identifying unusual anatomical variants of the LA and PVs. Finally, MSCT and anatomical mapping were performed in different times; changes in patient’s status meanwhile could potentially affect the comparison between MSCT images and real-time anatomical reconstruction. Nevertheless, in order to minimize this potential limitation, care was taken to perform anatomical mapping within 24 hours from MSCT and in the same heart rhythm as during MSCT. Conclusions Overall accuracy in the LA and PV anatomical reconstruction is superior with FAM compared to ultrasound-guided approach. FAM recreates a 3D anatomical reconstruction of the LA and PV that is highly comparable to preacquired MSCT. On the contrary, ultrasound-derived anatomical reconstruction gives a significant underestimate of the true dimensions of both LA and PV.
References 1. Pappone C, Rosanio S, Oreto G, Tocchi M, Gugliotta F, Vicedomini G, Salvati A, et al. Circumferential radiofrequency ablation of pulmonary vein ostia: A new anatomic approach for curing atrial fibrillation. Circulation 2000; 102:2619–2628. 2. Rossillo A, Indiani S, Bonso A, Themistoclakis S, Corrado A, Raviele A. Novel ICE-guided registration strategy for integration of electroanatomical mapping with three-dimensional CT/MR images to guide catheter ablation of atrial fibrillation. J Cardiovasc Electrophysiol 2009; 20:374–378. 3. Singh SM, Heist EK, Donaldson DM, Collins RM, Chevalier J, Mela T, Ruskin JN, et al. Image integration using intracardiac ultrasound to guide catheter ablation of atrial fibrillation. Heart Rhythm 2008; 5:1548–1555. 4. den Uijl DW, Tops LF, Tolosana JM, Schuijf JD, Trines SA, Zeppenfeld K, Bax JJ, et al. Real-time integration of intracardiac echocardiography and multislice computed tomography to guide radiofrequency catheter ablation for atrial fibrillation. Heart Rhythm 2008; 5:1403–1410. 5. Pratola C, Baldo E, Artale P, Marcantoni L, Toselli T, Percoco G, Sassone B, et al. Different image integration modalities to guide AF
194
ablation: Impact on procedural and fluoroscopy times. Pacing Clin Electrophysiol 2011; 34:422–430. 6. Okumura Y, Watanabe I, Kofune M, Nagashima K, Sonoda K, Mano H, Ohkubo K, et al. Effect of catheter tip-tissue surface contact on three-dimensional left atrial and pulmonary vein geometries: Potential anatomic distortion of 3D ultrasound, fast anatomical mapping, and merged 3D CT-derived images. J Cardiovasc Electrophysiol 2013; 24:259–266. 7. Schwartzman D, Zhong H. On the use of CartoSound for left atrial navigation. J Cardiovasc Electrophysiol 2010; 21:656–664. 8. Calkins H, Kuck KH, Cappato R, Brugada J, Camm AJ, Chen SA, Crijns HJ, et al. Heart Rhythm Society Task Force on Catheter and Surgical Ablation of Atrial Fibrillation. 2012 HRS/EHRA/ECAS expert consensus statement on catheter and surgical ablation of atrial fibrillation: Recommendations for patient selection, procedural techniques, patient management and followup, definitions, endpoints, and research trial design: A report of the Heart Rhythm Society (HRS) Task Force on Catheter and Surgical Ablation of Atrial Fibrillation. Developed in partnership with the European Heart Rhythm Association (EHRA), a registered branch
February 2015
PACE, Vol. 38
ANATOMICAL MAPPING FOR ATRIAL FIBRILLATION ABLATION of the European Society of Cardiology (ESC) and the European Cardiac Arrhythmia Society (ECAS); and in collaboration with the American College of Cardiology (ACC), American Heart Association (AHA), the Asia Pacific Heart Rhythm Society (APHRS), and the Society of Thoracic Surgeons (STS). Endorsed by the governing bodies of the American College of Cardiology Foundation, the American Heart Association, the European Cardiac Arrhythmia Society, the European Heart Rhythm Association, the Society of Thoracic Surgeons, the Asia Pacific Heart Rhythm Society, and the Heart Rhythm Society. Heart Rhythm 2012; 9:632–696. 9. Tops LF, Bax JJ, Zeppenfeld K, Jongbloed MR, Lamb HJ, van der Wall EE, Schalij MJ. Fusion of multislice computed tomography imaging with three-dimensional electroanatomic mapping to guide radiofrequency catheter ablation procedures. Heart Rhythm 2005; 10:1076–1081. 10. Tops LF, Schalij MJ, den Uijl DW, Abraham TP, Calkins H, Bax JJ. Image integration in catheter ablation of atrial fibrillation. Europace 2008; 10(Suppl 3):iii48–56. 11. Okumura Y, Henz BD, Johnson SB, Bunch TJ, O’Brien CJ, Hodge DO, Altman A, et al. Three-dimensional ultrasound for image-guided mapping and intervention: Methods, quantitative validation, and
12.
13.
14.
15.
clinical feasibility of a novel multimodality image mapping system. Circ Arrhythm Electrophysiol 2008; 1:110–119. Kuck KH, Reddy VY, Schmidt B, Natale A, Neuzil P, Saoudi N, Kautzner J, et al. A novel radiofrequency ablation catheter using contact force sensing: Toccata study. Heart Rhythm 2012; 9: 18–23. Ozcan EE, Szeplaki G, Tahin T, Osztheimer I, Szilagyi S, Apor A, Horvath PM, et al. Impact of respiration gating on image integration guided atrial fibrillation ablation. Clin Res Cardiol 2014; 103:723– 731 Jongbloed MR, Bax JJ, Lamb HJ, Dirksen MS, Zeppenfeld K, van der Wall EE, de Ross A, et al. Multislice computed tomography versus intracardiac echocardiography to evaluate the pulmonary veins before radiofrequency catheter ablation of atrial fibrillation: A head-to-head comparison. J Am Coll Cardiol 2005; 45:343– 350. Bertaglia E, Della Bella P, Tondo C, Proclemer A, Bottoni N, De Ponti R, Landolina M, et al. Image integration increases efficacy of paroxysmal atrial fibrillation catheter ablation: Results from the CartoMergeTM Italian Registry. Europace 2009; 11:1004– 1010.
Supporting Information Additional Supporting Information may be found in the online version of this article at the publisher’s website: Figure 1: Intermethods agreement for LA and PV diameters measured with ICE versus MSCT (upper panels) and with FAM versus MSCT (lower panels). Figure 2: Interoperator variability for LA and PV diameters measured with ICE (upper panels), FAM (middle panels), and MSCT (lower panels).
PACE, Vol. 38
February 2015
195
This document is a scanned copy of a printed document. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material.
Background: Accuracy in left atrial (LA) anatomical reconstruction is crucial to the safe and effective performance of catheter ablation of atrial fibrillation (AF). The aim of this study was to evaluate the accuracy of LA reconstruction performed with intracardiac echocardiography (ICE) as compared to fast anatomical mapping (FAM) both integrated in the CARTO mapping system (Biosense Webster, Diamond Bar, CA, USA). Methods: A multislice computed tomography (MSCT) was preacquired from 29 patients with AF who underwent catheter ablation and 3D-LA geometry was reconstructed using both ICE and FAM separately. The accuracy of the LA anatomical definition was evaluated by comparing LA volumes, LA and pulmonary vein (PV) diameters obtained using ICE and FAM versus MSCT (gold standard). Results: Anterior-posterior and superior-inferior LA diameters were shorter in ICE versus MSCT (32 ± 10 vs 46 ± 9 mm and 48 ± 7 vs 53 ± 7 mm, P < 0.01) but similar in FAM versus MSCT (45 ± 9 vs 46 ± 9 mm and 52 ± 10 vs 53 ± 7 mm). Latero-septal LA diameter was similar in ICE versus MSCT (63 ± 11 vs 63 ± 9 mm) but larger in FAM versus MSCT (69 ± 9 vs 63 ± 9 mm, P < 0.001). LA volume was lower in ICE versus MSCT (73 ± 30 mL vs 116 ± 45 mL, P < 0.0001) and slightly larger in FAM versus MSCT (132 ± 45 vs 116 ± 45 mL, P = 0.06). PV diameters were similar in FAM versus MSCT but significantly underestimated with ICE. Conclusions: Overall accuracy in the LA and PV anatomical reconstruction was found to be superior with FAM compared to ICE-guided approach. ICE resulted in a significant underestimate of both LA and PV dimensions, while FAM slightly overestimated LA geometry. (PACE 2015; 38:187–195) atrial fibrillation ablation, intracardiac echocardiography, fast anatomical mapping, CARTO system
Introduction Left atrial (LA) three-dimensional (3D) reconstruction is essential for the performance of safe and effective radiofrequency (RF) catheter ablation of atrial fibrillation (AF). Different techniques have been proposed for recreating LA anatomy in order to guide AF ablation: point-by-point electroanatomical mapping,1 intracardiac echocardiography (ICE),2,3 and fast anatomical mapping (FAM). Specifically, these two latter technologies have recently been developed and integrated
Address for reprints: Roberto Rordorf, M.D., Department of Cardiology, IRCCS Fondazione Policlinico S. Matteo, P.le Golgi 19, 27100 Pavia, Italy. Fax: 39-0382-503161; e-mail: [email protected] Received February 25, 2014; revised September 21, 2014; accepted September 24, 2014. doi: 10.1111/pace.12539
R R into the CARTO 3 System with CARTOMERGE Module (Biosense Webster, Diamond Bar, CA, USA). The ultrasound-based 3D imaging system R (CARTOSOUND Module, Biosense Webster)4,5 allows to recreate LA and pulmonary vein (PV) derived from 2D ICE images obtained using an ICE catheter placed in the right atrium (RA). FAM allows a quick recreation of 3D chamber geometries by moving sensor-based catheters placed in the left atrium. Both these methods can be used alone or after integration with a preacquired computer tomography (CT) or magnetic resonance (MR) image of the LA and PV and have the advantage of limiting x-ray exposure. Nowadays, few data comparing these two real-time reconstruction strategies are available.6 The primary aim of this study was to compare the accuracy in overall reconstruction of LA and PV geometries obtained with ICE versus FAM, considering CT images as the gold standard. A mapping technique that
©2014 Wiley Periodicals, Inc. PACE, Vol. 38
February 2015
187
RORDORF, ET AL.
Table I. Baseline Characteristics of the Study Population Age (years) Gender (M/F) Smoking Systemic hypertension Diabetes Dyslipidemia AF (paroxysmal/persistent) AF onset-AF ablation time (months) Number of antiarrhythmic drugs before ablation 1 drug 2 drugs 3 drugs LV ejection fraction (< 50%) Left atrial diameter (mm)
56 ± 9 28/1 14 (48%) 17 (58%) 2 (10%) 12 (41%) 16/13 88 ± 74 7 (24%) 9 (31%) 13 (45%) 6 (21%) 45 ± 6
AF = atrial fibrillation; LV = left ventricular.
allows to recreate a precise anatomy of both the LA and PV could potentially be used as a stand-alone tool.7 Therefore the primary objective of our study tried to address the issue whether the mapping techniques under study could be used without the need for preacquired CT or MR images. The secondary objective of the study was to evaluate the accuracy of ICE versus FAM in the integration process with CT images. Methods Study Population Twenty-nine consecutive patients (28 males, age 56 ± 9 years) with symptomatic drugrefractory AF underwent RF catheter ablation at the Electrophysiologic Unit of the Department of Cardiology and were enrolled in the study. The patient population had paroxysmal or persistent AF refractory to at least one antiarrhythmic drug. Before the procedure both transthoracic and transesophageal echocardiographic evaluations were performed. LA dimension and left ventricular ejection fraction were estimated and the presence of thrombi within the cardiac chambers was excluded. Oral anticoagulation was interrupted 2 days before hospital admittance and intravenous heparin was administered during the procedure to maintain activated clotting time > 300 ms.8 The study protocol was approved by the institutional ethical committee and written informed consent was obtained from all patients. Baseline characteristics of the study population are summarized in Table I. Multislice Computed Tomography All patients underwent multislice CT (MSCT) to reconstruct PV and LA anatomy. MSCT was
188
obtained on a Siemens Somatom Definition with 64-slice dual-source configuration (Siemens AG, Healthcare Sector, Erlangen, Germany) within 24 hours before the ablation procedure to avoid significant modifications in LA anatomy and volume secondary to changes in preload and afterload conditions. Retrospectively electrocardiogram-gated spiral scanning was performed with a collimation of 1.2 mm during injection of 80 mL of contrast medium (Iomeron 400 mg I/mL, Bracco, Milan, Italy) at a flow rate of 4 mL/s. Cardiac telediastolic phase was automatically determined in patients with sinus rhythm, while manual editing was performed on patients with AF to minimize the motion artifact. Images were acquired during the endexpiratory phase. Multiplanar reconstruction, maximum intensity projection, and volume rendering were used to define and describe LA anatomy. LA volume was measured with a dedicated volume analysis software (Syngo VE32B Volume, software version 1.0, 2008, Siemens AG). The LA was manually outlined on axial images by using a 3-mm thickness sequence; afterwards the contours were semiautomatically determined and volume was calculated. Ostial PV diameters, LA anterior-posterior, lateralseptal, and superior-inferior diameters were also manually determined. MSCT images (slice thickness 1.5 mm, position increment 0.7 mm) were then stored on a CD and data were loaded into the image integration module of the R mapping system (CARTOMERGE Module, Biosense Webster). Afterwards a segmentation process of MSCT images was performed as previously described.9,10 3D reconstructions of LA and PVs were isolated from the surrounding structures and exported to the mapping system for the subsequent mapping and integration processes. LA and PV anatomy obtained with MSCT was considered the gold standard and used to evaluate the accuracy of anatomical reconstruction obtained with ICE and FAM. Anatomical Reconstruction with ICE After vascular access was obtained, a 10-Fr R ICE catheter (SOUNDSTAR Catheter, 64-element, 5.5–10.0 MHz, Biosense Webster) was placed in the left femoral vein and advanced into the RA. The ICE catheter tip incorporates a sensor to establish its location and direction and to show the ultrasound beam projection in the mapping system. Sequential clockwise manipulation of the ICE probe from the RA allowed for visualization of the LA, LA appendage (LAA), and PVs. 2D ultrasound segment images were acquired during
February 2015
PACE, Vol. 38
ANATOMICAL MAPPING FOR ATRIAL FIBRILLATION ABLATION
end-expiratory phase and gated to coronary sinus atrial electrogram in patients in sinus rhythm and to 80% of the R-R interval in patients with AF. ICE was performed using a Sequoia Cypress ultrasound system (Siemens Medical Solutions USA, Mountain View, CA, USA), which transferred real-time ICE data to the mapping system. The endocardial surface contours of each ultrasound image were manually drawn. All contours were used to automatically generate a single 3D anatomical map that included both LA R and PVs with the use of the CARTOSOUND Module (Biosense Webster). The volume of the chambers R was automatically provided by the CARTOSOUND Module as volume subtended to the interpolated map surface. Anatomical Reconstruction with FAM After double-transseptal catheterization, a 3.5-mm cool saline-irrigated ablation catheter R R (NAVISTAR THERMOCOOL Catheter, Biosense Webster) and a multipolar circular catheter R (LASSO NAV Catheter, Biosense Webster), both R compatible with CARTO mapping system, were advanced in the LA and PVs and used for anatomical reconstruction. High-resolution 3D chamber volumes were created from continuous recording of the catheter movements within the cardiac chambers to outline the structures. Anatomical rendering was obtained by the mapping system in a stable mode calculating the average of the catheter positions in the running second. The mapping catheter was placed 1 cm inside each PV and multiple positions were acquired while the catheter was rotated inside the vein and pulled back toward the ostium. The highest definition setting was used to better separate each vein and the lateral veins from the LAA. A 3D shell of the LA and LAA was created roving R R R the NAVISTAR , THERMOCOOL , or LASSO NAV catheters into the atrium: a lower definition setting was applied for the LA to obtain sufficiently fast and detailed atrial reconstruction. During volume sampling, catheter movements during inspiration phase were manually excluded. The R CARTO system automatically calculates the volume of the chambers as the volume subtended to the interpolated map surface. Electroanatomical points were not acquired and adapted to LA surface during PV and LA reconstruction with FAM. Anatomical reconstruction of the LA was performed both with ICE and FAM during the same cardiac rhythm (either sinus rhythm or AF) as during the acquisition of MSCT images in all patients. Anatomical mapping was performed in sinus rhythm in 16 patients and in AF in 13 patients.
PACE, Vol. 38
Accuracy of Overall Anatomical Reconstruction of LA and PVs In order to establish the overall accuracy of 3D anatomical reconstruction of LA and PVs, obtained with ICE versus FAM, MSCT images were considered the gold standard. PV, LA diameters, and LA volumes were measured with ICE and FAM and compared to those obtained with MSCT. The anterior-posterior, latero-septal, and superiorinferior diameters of the LA were assessed with MSCT as in Figure 1 and compared to those obtained with both ICE and FAM reconstruction. Figure 2 shows an example of the superior-inferior LA diameter calculated with MSCT, FAM, and ICE in the same patient. In addition, the overall LA volume resulting from ICE and FAM was compared to that obtained with MSCT. In the computation of the overall LA volume, care was taken to include the proximal part (1 cm inside) of each PV. The ability of each mapping technique to identify the number of PVs per patient and the presence of separate versus common ostia were also assessed. Finally, PV diameters were measured with each technique. PV diameters on ICE were measured at the widest point; PV diameters on MSCT and FAM were measured at the same angulation as with ICE to obtain comparable data. LA and PV diameters were measured by two operators independently and interoperator variability was assessed. The time needed for LA reconstruction with ICE versus FAM was calculated and x-ray exposure time was also analyzed.
Accuracy of the Registration Process with MSCT After creating an anatomical reconstruction with ICE and FAM, each map was integrated with MSCT separately. First of all, the ICE anatomical map was displayed adjacent to the MSCT image. A landmark point was placed at a distinct anatomical structure (usually at the PV ostia) on the ICE map and at the corresponding point on the MSCT image. Visual alignment and surface registration was then performed as described previously.4 The FAM map was then displayed adjacent to the MSCT image and the same integration process was performed for the FAM anatomical map, without considering the ICE contours. A new landmark point was identified on the FAM map and on the corresponding point on the MSCT, and a new visual alignment and surface registration process were performed. The accuracy of the registration process was assessed by evaluating the mean distance and the standard deviation of the distances between the MSCT surface and each point on the ICE and FAM map separately.
February 2015
189
RORDORF, ET AL.
Figure 1. Left atrial diameters. Left panel: The anterior-posterior diameter (sagittal diameter) of the left atrium calculated as the distance between the central part of the posterior wall to the central part of the mitral annulus. Middle panel: The superior-inferior diameter (longitudinal diameter) calculated as the distance between the central part of the roof to the central part of the inferior wall. Right panel: The latero-septal diameter (transverse diameter) calculated as the distance between the central part of the lateral carina to the central part of the septal carina.
Figure 2. Right lateral view showing LA superior-inferior diameter calculated with FAM (left), MSCT (middle), and ICE (right) in the same patient. FAM = fast anatomical mapping; ICE = intracardiac echocardiography; LA = left atrial; MSCT = multislice computed tomography.
Statistical Analysis Data are presented as mean ± standard deviation for continuous variables and as count and percentage for categorical variables. Statistical comparisons were performed with the two-tailed paired Student’s t-test. Bland and Altman limits of agreement and Lin’s concordance correlation coefficient (Lin R), with 95% confidence intervals (95% confidence interval) were computed to assess interoperators and intermethods agreement. P < 0.05 was considered statistically significant. The Bonferroni correction was applied when comparing each of the two methods against MSCT (P < 0.025 for significance). Statistical analysis was performed with MedCalc (MedCalc Software,
190
Ostend, Belgium) and Stata 13 (StataCorp, College Station, TX, USA). Results The time required for a comprehensive LA and PV anatomical reconstruction was similar for ICE and FAM techniques (33 ± 14 minutes vs 30 ± 12 minutes, P = 0.410). Fluoroscopy time needed for ICE reconstruction was significantly lower than the time required for FAM reconstruction (2.1 ± 4.4 minutes vs 8.9 ± 4.6 minutes, P < 0.0001). A mean of 23 ± 10 ultrasound contours were used to obtain a global anatomical rendering of the LA and PVs. There were no major complications related to the ablation procedure or to the protocol.
February 2015
PACE, Vol. 38
ANATOMICAL MAPPING FOR ATRIAL FIBRILLATION ABLATION
Accuracy of Overall LA and PV Anatomical Reconstruction The ability of ICE anatomical mapping to recreate a comprehensive LA 3D shell resulted suboptimal due mainly to an underestimate of the LA dimension in the sagittal and longitudinal diameter. Indeed, compared to the gold-standard MSCT, ICE-derived anteriorposterior and superior-inferior diameters resulted significantly shorter (32 ± 10 mm vs 46 ± 9 mm, P < 0.0001, 48 ± 7 mm vs 53 ± 7 mm, P = 0.001). Conversely, the latero-septal LA diameter obtained with ICE reconstruction was comparable to MSCT (63 ± 11 mm vs 63 ± 9 mm, P = 0.956). On the contrary, FAM-derived 3D LA reconstruction was more similar to that obtained with MSCT except for a significant overestimation of the LA dimension in the transverse diameter. The latero-septal LA diameter was significantly larger with FAM compared to MSCT (69 ± 9 mm vs 63 ± 9 mm, P ࣘ 0.001). Both anterior-posterior and superior-inferior LA diameters were similar in FAM compared to MSCT (45 ± 9 mm vs 46 ± 9 mm, P = 0.107, 52 ± 10 mm vs 53 ± 7 mm, P = 0.511). Figure 3 summarizes the LA diameters obtained with ICE and FAM compared to MSCT. Similar results were found when the analysis was repeated separately in patients performing anatomical mapping in sinus rhythm and AF (see Table II). Overall LA volume was significantly underestimated by ICE compared to MSCT (73 ± 30 mL vs 116 ± 45 mL, P < 0.0001) and slightly overestimated by FAM versus MSCT (132 ± 45 mL vs 116 ± 45 mL, P = 0.06). Regarding the accuracy of ICE 3D reconstruction, the analysis was repeated after excluding those patients with less than 20 ICE contours available. Similar to the general population, ICE-derived anteriorposterior, superior-inferior diameters, and overall LA volume were significantly shorter compared to MSCT (32 ± 9 mm vs 46 ± 10 mm, P < 0.0001, 48 ± 7 mm vs 53 ± 10 mm, P = 0.001, 77 ± 31 mL vs 120 ± 47 mL, P < 0.0001). All patients showed four main PVs. FAM allowed to correctly identify all PVs in the entire population. On the contrary, ICE did not correctly reconstruct all PVs in five patients: in one patient a right inferior PV was not identified and in four patients the quality of the anatomical rendering was considered unsatisfactory (three right veins and one left vein). In 24 of 29 patients, the four PVs showed separated ostia. In five patients a common ostium was detectable (four left and one right). FAM allowed to identify a common ostium in all cases while in one case ICE identified
PACE, Vol. 38
the two veins but not the proximal part of the common ostium. Diameters of all four PVs were underestimated with ICE versus MSCT. On the contrary, PV diameters obtained with FAM were comparable to MSCT. Figure 4 summarizes the PV diameters obtained with the different mapping techniques. PV diameters obtained with ICE were still shorter compared to MSCT also when the analysis was limited to those patients with 20 or more acquired ICE contours. A better accuracy in overall anatomical reconstruction with FAM versus ICE was also confirmed by intermethods agreement analysis. Indeed, overall agreement with preacquired MSCT was superior with FAM for both LA and PV diameters (see Fig. S1). Interoperator variability was excellent for PV and LA diameters measurements with all methods (see Fig. S2). Accuracy of the Registration Process of ICE and FAM Maps with MSCT To enable a registration process with preacquired MSCT images, a fiducial landmark point was identified on each recreated anatomical map. A landmark point at the PV ostia was used in the majority of the cases in both FAM and ICE reconstructions. Specifically, a landmark point was identified at a PV ostium in 19 cases and at the PV carina in 10 cases with FAM. Similarly, with ICE a landmark point was identified at a PV ostium in 19 cases, at the PV carina in six cases, and within the LA in four cases. After each separate registration process, the accuracy of the map to MSCT merge was superior with ICE compared to FAM. Actually, both the mean distance and the standard deviation of the distances between the MSCT surface and each point on the reconstructed anatomical map were significantly lower for ICE compared to FAM (2.5 ± 0.5 mm vs 3.6 ± 0.7 mm; 1.8 ± 0.4 mm vs 2.9 ± 0.8 mm, P < 0.0001 for both). Displacement of the landmark point from its initial location after the merge process was similar for ICE versus FAM (8 ± 5 mm vs 8 ± 4 mm, P = 0.914). Discussion Our study reports a head-to-head comparison of two techniques for LA and PV anatomical reconstruction to guide AF ablation and highlights important differences between ICE-guided as opposed to FAM 3D rendering, with significant clinical implications for physicians performing ablation. We found that 3D ultrasound-derived anatomical reconstruction yields a significant underestimate of true LA volume, in agreement with
February 2015
191
RORDORF, ET AL.
Figure 3. Comparison of LA diameters obtained with ICE and FAM versus MSCT. AP = anterior-posterior; SI = superior-inferior; LS = latero-septal; ICE = intracardiac echocardiography, MSCT = multislice computer tomography; FAM = fast anatomical mapping. Table II. Left Atrial Diameters in Patients Performing Anatomical Mapping in Sinus Rhythm and in Atrial Fibrillation Anterior-Posterior Diameter (mm)
Patients in SR (n = 16) Patients in AF (n = 13)
Superior-Inferior Diameter (mm)
Latero-Septal Diameter (mm)
ICE
MSCT
FAM
ICE
MSCT
FAM
ICE
MSCT
FAM
32 ± 9* 30 ± 12*
44 ± 8 50 ± 9
43 ± 8 47 ± 9
48 ± 7 48 ± 6*
52 ± 8 53 ± 5
52 ± 10 51 ± 11
61 ± 12 67 ± 7
59 ± 5 66 ± 8
68 ± 10** 71 ± 6
*P < 0.01 for ICE versus MSCT; **P < 0.01 for FAM versus MSCT.
previous data from the literature.6,7,11 However, the analysis of three different LA diameters in our study gives a better understanding of the limitations of ICE in recreating a comprehensive anatomical reconstruction as described in Figure 3. Actually, our data highlight that the underestimate of LA volume is especially due to a significant misjudgment of the true measurements of the longitudinal and sagittal diameters. The use of an ICE catheter placed in the RA allows a clear definition of the transversal extension of the LA, through a simple clockwise rotation of the catheter from lateral to more septal views. Further manipulation and angulation of the catheter allows reconstruction of a 3D rendering of the LA, but still with a certain degree of underestimate of the true LA dimensions, probably due to a more challenging definition of the anterior, superior, and inferior borders of the LA. This limitation could potentially be overcome by placing the ICE catheter in the coronary sinus, in the right ventricle or directly in the LA as previously described,3 but a more extensive approach is time consuming and could increase complications. The aim of our study was to assess the accuracy of ultrasound LA reconstruction using a simple approach and taking a reasonable time. Therefore,
192
we limited the ultrasound exploration by placing the ICE catheter in the RA only. Our study assessed the accuracy of FAM for recreating a comprehensive 3D anatomical reconstruction of the LA. We demonstrated that FAM allows generation of a 3D shell of the LA that is very similar to that obtained with MSCT. It is worth noting that FAM-derived longitudinal and sagittal diameters almost matched LA measurements resulting from MSCT. The transverse diameter was the only LA dimension that was significantly overestimated by FAM. This could be due to atrial wall deformation caused by catheter manipulation: the lateral carina is directly opposite the transseptal access and this could explain why greater forces are usually applied to the mapping catheter in this area. This limitation could be probably overcome by the use of contact force sensing catheters.12 Another possible explanation for latero-septal diameter overestimate is that the FAM algorithm could be affected by higher interpolation between adjacent veins in the carina, despite the use of a highdefinition setting in this area. A slight overestimation of LA dimensions with FAM compared to MSCT could also be explained by the fact that we used a CARTO 3
February 2015
PACE, Vol. 38
ANATOMICAL MAPPING FOR ATRIAL FIBRILLATION ABLATION
Figure 4. Comparison of PV diameters obtained with ICE and FAM versus MSCT. FAM = fast anatomical mapping; ICE = intracardiac echocardiography; LIPV = left inferior pulmonary vein; LSPV = left superior pulmonary vein; MSCT = multislice computer tomography; PV = pulmonary vein; RIPV = right inferior pulmonary vein; RSPV = right superior pulmonary vein.
software version that does not allow respiratory gating. Even better accuracy in 3D rendering of the LA is to be expected with the more recent CARTO 3 version embedded with the respiratory gating module.13 Another important finding of our study is that FAM was more reliable compared to ICE in rendering the anatomy of the PVs. In our experience FAM was able to correctly identify all PVs, including five cases with common ostia, while ICE was considered unsatisfactory in a few cases. Moreover, FAM was more consistent with MSCT in PV diameter definition. Actually, the PV diameters obtained with FAM nearly matched those derived from MSCT. On the contrary, ICE-derived PV diameters were significantly smaller compared to those obtained with CT. This is in agreement with data from other groups.4,14 Left PVs are usually visualized with ICE in longitudinal cross-section, potentially missing the middle and widest part of the vein. Regarding
PACE, Vol. 38
right-sided PVs, these are usually visualized in transversal cross-section and the underestimate of the true diameter could be due to the more challenging visualization of the exact PV-LA junction. The second aim of our study was to evaluate the accuracy of the integration (Merge) process15 with MSCT of the two mapping techniques under investigation. ICE-derived anatomy yielded a better integration with MSCT compared to FAM. This finding might appear counterintuitive to the aforementioned results of a more comprehensive overall 3D LA rendering obtained with FAM. Nevertheless, compared to FAM, ICE is likely to provide a more accurate direct visualization of the fiducial points used for a more reliable superimposition with MSCT. Although ICE was probably insufficient for correctly visualizing some areas of the anterior part of the LA, it was able to define the posterior wall and the posterior PV-LA junction very precisely to allow a good
February 2015
193
RORDORF, ET AL.
alignment with MSCT. Another possible explanation is that the FAM-derived anatomical map is made up of a greater number of “virtual points” compared to ICE, given by the continuous acquisition during catheter manipulation. The greater “virtual point” density in FAM versus ICE increases the chances of mismatch between the MSCT surface and the recreated anatomical map. Our finding of a better integration accuracy with ICE versus FAM should be, however, interpreted with caution as it is based only on the overall mean distance between the MSCT surface and the recreated anatomical maps. Nonetheless, our study corroborates previous data suggesting the utility of ICE for integration processes.4 Clinical Implications A variety of issues are taken into account by electrophysiologists when choosing the preferred mapping technique for performing AF ablation: procedure and fluoroscopy time, the time spent in the left chambers, cost efficacy, and safety. The ideal mapping technique should be quick, reduce both fluoroscopy time and the time spent in the left chamber, and provide accurate anatomical information that could potentially replace MSCT. Our data add some important insights and illustrate the advantages and drawbacks of each technique. ICE-derived mapping techniques reduce both fluoroscopy time and the time spent in the left atrium, but provide an unsatisfactory overall 3D rendering, at least in a relative short procedural time, making this technique unsuitable in our opinion for replacing MSCT. Nonetheless, the ICE direct visualization of distinct anatomical structures allows a good alignment with MSCT and makes this technique suitable for integration processes. On the other hand, in our experience FAM allowed a more accurate rendering of both
LA and PV, which could potentially replace the need for preacquired MSCT, with a reduction in both procedural expenses and radiation exposure. Limitations The data presented do not allow a comprehensive analysis of the safety and efficacy issues of ICE and FAM for performing AF ablation. Our study was limited to a head-to-head investigation of the two anatomical mapping techniques, with the advantage of performing the comparison on the same patient. In our population, we did not find supernumerary veins or LA diverticula. This is probably due to the small sample size of our study. A larger study would be required to analyze the accuracy of ultrasound and FAM mapping in identifying unusual anatomical variants of the LA and PVs. Finally, MSCT and anatomical mapping were performed in different times; changes in patient’s status meanwhile could potentially affect the comparison between MSCT images and real-time anatomical reconstruction. Nevertheless, in order to minimize this potential limitation, care was taken to perform anatomical mapping within 24 hours from MSCT and in the same heart rhythm as during MSCT. Conclusions Overall accuracy in the LA and PV anatomical reconstruction is superior with FAM compared to ultrasound-guided approach. FAM recreates a 3D anatomical reconstruction of the LA and PV that is highly comparable to preacquired MSCT. On the contrary, ultrasound-derived anatomical reconstruction gives a significant underestimate of the true dimensions of both LA and PV.
References 1. Pappone C, Rosanio S, Oreto G, Tocchi M, Gugliotta F, Vicedomini G, Salvati A, et al. Circumferential radiofrequency ablation of pulmonary vein ostia: A new anatomic approach for curing atrial fibrillation. Circulation 2000; 102:2619–2628. 2. Rossillo A, Indiani S, Bonso A, Themistoclakis S, Corrado A, Raviele A. Novel ICE-guided registration strategy for integration of electroanatomical mapping with three-dimensional CT/MR images to guide catheter ablation of atrial fibrillation. J Cardiovasc Electrophysiol 2009; 20:374–378. 3. Singh SM, Heist EK, Donaldson DM, Collins RM, Chevalier J, Mela T, Ruskin JN, et al. Image integration using intracardiac ultrasound to guide catheter ablation of atrial fibrillation. Heart Rhythm 2008; 5:1548–1555. 4. den Uijl DW, Tops LF, Tolosana JM, Schuijf JD, Trines SA, Zeppenfeld K, Bax JJ, et al. Real-time integration of intracardiac echocardiography and multislice computed tomography to guide radiofrequency catheter ablation for atrial fibrillation. Heart Rhythm 2008; 5:1403–1410. 5. Pratola C, Baldo E, Artale P, Marcantoni L, Toselli T, Percoco G, Sassone B, et al. Different image integration modalities to guide AF
194
ablation: Impact on procedural and fluoroscopy times. Pacing Clin Electrophysiol 2011; 34:422–430. 6. Okumura Y, Watanabe I, Kofune M, Nagashima K, Sonoda K, Mano H, Ohkubo K, et al. Effect of catheter tip-tissue surface contact on three-dimensional left atrial and pulmonary vein geometries: Potential anatomic distortion of 3D ultrasound, fast anatomical mapping, and merged 3D CT-derived images. J Cardiovasc Electrophysiol 2013; 24:259–266. 7. Schwartzman D, Zhong H. On the use of CartoSound for left atrial navigation. J Cardiovasc Electrophysiol 2010; 21:656–664. 8. Calkins H, Kuck KH, Cappato R, Brugada J, Camm AJ, Chen SA, Crijns HJ, et al. Heart Rhythm Society Task Force on Catheter and Surgical Ablation of Atrial Fibrillation. 2012 HRS/EHRA/ECAS expert consensus statement on catheter and surgical ablation of atrial fibrillation: Recommendations for patient selection, procedural techniques, patient management and followup, definitions, endpoints, and research trial design: A report of the Heart Rhythm Society (HRS) Task Force on Catheter and Surgical Ablation of Atrial Fibrillation. Developed in partnership with the European Heart Rhythm Association (EHRA), a registered branch
February 2015
PACE, Vol. 38
ANATOMICAL MAPPING FOR ATRIAL FIBRILLATION ABLATION of the European Society of Cardiology (ESC) and the European Cardiac Arrhythmia Society (ECAS); and in collaboration with the American College of Cardiology (ACC), American Heart Association (AHA), the Asia Pacific Heart Rhythm Society (APHRS), and the Society of Thoracic Surgeons (STS). Endorsed by the governing bodies of the American College of Cardiology Foundation, the American Heart Association, the European Cardiac Arrhythmia Society, the European Heart Rhythm Association, the Society of Thoracic Surgeons, the Asia Pacific Heart Rhythm Society, and the Heart Rhythm Society. Heart Rhythm 2012; 9:632–696. 9. Tops LF, Bax JJ, Zeppenfeld K, Jongbloed MR, Lamb HJ, van der Wall EE, Schalij MJ. Fusion of multislice computed tomography imaging with three-dimensional electroanatomic mapping to guide radiofrequency catheter ablation procedures. Heart Rhythm 2005; 10:1076–1081. 10. Tops LF, Schalij MJ, den Uijl DW, Abraham TP, Calkins H, Bax JJ. Image integration in catheter ablation of atrial fibrillation. Europace 2008; 10(Suppl 3):iii48–56. 11. Okumura Y, Henz BD, Johnson SB, Bunch TJ, O’Brien CJ, Hodge DO, Altman A, et al. Three-dimensional ultrasound for image-guided mapping and intervention: Methods, quantitative validation, and
12.
13.
14.
15.
clinical feasibility of a novel multimodality image mapping system. Circ Arrhythm Electrophysiol 2008; 1:110–119. Kuck KH, Reddy VY, Schmidt B, Natale A, Neuzil P, Saoudi N, Kautzner J, et al. A novel radiofrequency ablation catheter using contact force sensing: Toccata study. Heart Rhythm 2012; 9: 18–23. Ozcan EE, Szeplaki G, Tahin T, Osztheimer I, Szilagyi S, Apor A, Horvath PM, et al. Impact of respiration gating on image integration guided atrial fibrillation ablation. Clin Res Cardiol 2014; 103:723– 731 Jongbloed MR, Bax JJ, Lamb HJ, Dirksen MS, Zeppenfeld K, van der Wall EE, de Ross A, et al. Multislice computed tomography versus intracardiac echocardiography to evaluate the pulmonary veins before radiofrequency catheter ablation of atrial fibrillation: A head-to-head comparison. J Am Coll Cardiol 2005; 45:343– 350. Bertaglia E, Della Bella P, Tondo C, Proclemer A, Bottoni N, De Ponti R, Landolina M, et al. Image integration increases efficacy of paroxysmal atrial fibrillation catheter ablation: Results from the CartoMergeTM Italian Registry. Europace 2009; 11:1004– 1010.
Supporting Information Additional Supporting Information may be found in the online version of this article at the publisher’s website: Figure 1: Intermethods agreement for LA and PV diameters measured with ICE versus MSCT (upper panels) and with FAM versus MSCT (lower panels). Figure 2: Interoperator variability for LA and PV diameters measured with ICE (upper panels), FAM (middle panels), and MSCT (lower panels).
PACE, Vol. 38
February 2015
195
This document is a scanned copy of a printed document. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material.
